Disney, Universal, and Warner Bros. Copyright Claims Against Chinese AI Developer Hailuo Advance After US Federal Judge Refuses to Dismiss
A US federal judge in the Central District of California has refused to dismiss copyright infringement claims brought by Walt Disney, Universal, and Warner Bros. — alongside their subsidiaries — against the makers of Hailuo AI, an artificial intelligence video generation programme developed by Shanghai Xiyu Jizhi Technology Co. (trading as MiniMax) and Singapore-based Nanonoble Pte. The studios convinced the court that they had made plausible claims of both direct and secondary infringement (secondary infringement meaning liability for enabling or facilitating another's infringing acts, rather than committing them directly) arising from Hailuo AI's ability to generate near-perfect likenesses of the studios' well-known fictional characters. The refusal to dismiss at the pleadings stage means the case will proceed to discovery (the pre-trial evidence-gathering process) and potentially trial. The case is significant because it directly tests whether AI-generated output that closely mimics copyrighted characters constitutes infringement, and whether the developers of the underlying AI model bear secondary liability for that output. Both questions are live in multiple jurisdictions — the UK included, where CDPA 1988 (the Copyright, Designs and Patents Act 1988) and the emerging AI-and-copyright policy debate set a different but related statutory backdrop.
Why this matters
This ruling advances one of the most commercially important copyright questions in the AI era: whether generative AI models that reproduce the visual likeness of copyrighted characters infringe at the output stage, and whether model developers bear secondary liability. A finding of liability at trial would significantly reshape the risk calculus for generative AI developers globally, including those operating in the UK under CDPA 1988. The 'why now' is the rapid proliferation of video-generation AI tools capable of high-fidelity character reproduction, which has outpaced the licensing frameworks that studios rely on. For London law firms with IP and technology practices, the evolving US case law will be closely watched as a precursor to equivalent litigation before the UK's Intellectual Property Enterprise Court or the High Court.
On the Ground
A trainee on a copyright and AI dispute would assist with disclosure review and categorisation of technical evidence about the AI model's training data and output mechanisms, help prepare a chronology of the defendants' product development and the studios' rights ownership, and research skeleton argument points on the application of secondary infringement doctrine to AI-generated content.
Interview prep
Soundbite
Surviving dismissal means full discovery into how Hailuo AI was trained — which is where the real liability exposure lies.
Question you might get
“How does the doctrine of secondary copyright infringement apply to an AI developer whose model generates infringing output, and how might that analysis differ under UK law compared to US law?”
Full answer
A US federal court has allowed Disney, Universal, and Warner Bros.' copyright claims against Hailuo AI to proceed, finding the studios made plausible cases of both direct and secondary infringement from AI-generated character likenesses. For IP lawyers, the significance is that the case now moves into discovery, which will likely require the defendants to disclose their training data sets and model architecture — potentially revealing exactly which copyrighted materials the AI ingested. The wider picture is the global wave of studio-versus-AI-developer litigation that is establishing the foundational rules for AI-generated content liability across jurisdictions. This suggests that UK IP practices should be preparing parallel analysis under CDPA 1988, as similar claims will inevitably reach the English courts.
Sources
My notes
saved